It was a dark and stormy night. One of the largest hurricanes in recorded history curved through the Gulf of Mexicotoward southern states terrified and cowering at its approach, and crashed into Louisiana and Mississippi on Aug. 29, 2005, evoking a blackout of agonizing proportions.
New Orleans was submerged, thousands required rescue from the floods, and an untold number were drowned in the raging torrents brought by the storm. On the Thursday after the storm we began to hear, "The buses are coming." We were told to expect as many as 200 or more buses, each with 50 or more people rescued from New Orleans on board, some within hours.
Knowing we were already exceeding capacity at Reunion Arena, Dallas Fire Rescue and EMS medical direction expanded operations into the 140,000-square-foot parking lot beneath the Dallas Convention Center, where we built a home for nearly 2,000 people, with thousands of inflatable mattresses and various support services. And, with 24 hours to prepare, we crafted a medical operation that would ultimately see some 10,000+ patients from the 40,000+ evacuees who came to Dallas County from Hurricane Katrina and later from Rita. Indeed, the medical surge facility we built played a critical role in protecting local hospitals from being overwhelmed.
Why present a painful recollection of this disaster in an article about data management? Because at the Katrina Evacuee Center we found ourselves nearly overwhelmed by the new and ongoing medical needs of our guests from the coast: hypertension, heart disease, pulmonary illnesses, seizures off meds, OB/GYN conditions and advanced pregnancies, acute diseases superimposed upon chronic illnesses and, as important as any, the vast mental health needs of an urban population torn from its roots. Short of the often empty bottles of medications brought in their hands—often stained by Mississippi flood waters—we knew nothing of the medical history of most of these patients.
The need for a large personal health-care database became evident when, almost a week after the storm, a national pharmacy chain set up a formidable dispensary in a double-wide trailer in the entryway to the Convention Center. The pharmacy chain maintained a strong presence in Louisiana, and its databanks teemed with prescription health records of thousands of our evacuees. In the first full day of clinical operation at the Convention Center, we saw more than 1,100 patients in 24 hours, and this electronic database made our jobs a little easier and safer. But thousands of other patients weren’t in the database. Their paper records and complete medical histories were "water logged."
Not long after that episode, I took a break, cashed in some frequent flier miles and went to Italy. I stopped in Florence for a snack at a small café and at the end of the meal handed my credit card to the server. She whipped the card through a neat little wireless handheld checkout machine. In 10 seconds the machine had copied the card information, communicated with its local server, searched the database of the credit card company, approved the transaction and rolled out a receipt for me to sign. I couldn’t help but reflect that if we can be connected across the world from a financial data flow perspective, how is it that 10,000 evacuees from an adjoining state came mostly as absolute strangers from a medical data perspective? What an embarrassment—but also, what an opportunity!
Data ’R Us
Whether it’s global credit card purchases, driver’s licenses, criminal records or insurance policies, personal identification information surrounds us all in a veritable electron cloud.
EMS data flow is no exception. Nearly left behind in the history of medicine are handwritten patient care reports (PCRs). The average state of EMS systems today is an electronic data recording and management system that provides standard format data points to local servers that receive the data. Many state EMS agencies now require local EMS agencies to upload data points relative to clinical care for various public health purposes.
The 21st century saw the completion of the first national data-collecting standard for EMS: the National EMS Information System (NEMSIS), which brings together, publishes and supports a rigorous data-collection and management system for EMS. With that momentous achievement—and from an EMS systems administrative perspective it was indeed phenomenal —a road map appeared that would galvanize an industry into new ways of communicating.
Importance of Medical Data in EMS Systems
How could such an effort bring benefit to a local EMS system? A noteworthy example can be found in the Dallas Area BioTel EMS System. In Dallas County, a federation of cities was crafted 35 years ago; today, that federation is called the BioTel EMS System and comprises 17 EMS agencies ranging in size from 700+ paid paramedics responding in 65 vehicles to an agency that operates a single ambulance. These agencies cumulatively respond to about 250,000 EMS 9-1-1 calls per year in the Dallas metropolitan area, transporting to more than 20 emergency receiving facilities, including 17 hospitals with 24/7 PCI labs, as well as two Level I, one Level II and one Level III trauma centers.
When I first joined BioTel Medical Director Paul Pepe, MD, in Dallas in early 2001, we found that no fewer than eight discrete PCR databases had independently developed within the BioTel member cities. No common dataset existed. Data couldn’t be merged into a single dataset representing the system. To run a report as simple as "Destination hospital by age and gender of patients" required running eight separate reports, writing the results down by hand and then manually entering the data into a spreadsheet—a process that could take days.
Moreover, no standard format existed for data-tracking. In one database the administration of medications might be trackable by providers in separate, searchable fields. In another database that same data might be part of an "alpha" field and unrecoverable via a database search, which meant we couldn’t monitor narcotics administration across the EMS system, for example.
In addition, a serious flaw in the system appeared when the electronic data was transmitted to the BioTel EMS office. One large EMS system could send data updates only every six months or so, rendering the electronic data of little value as a continuous quality improvement (CQI) tool.
The early part of this decade saw the proliferation of dozens of vendors of electronic patient care reporting (ePCR) software. Two of our BioTel agencies were early users of different ePCR packages; however, neither could export data, and standard field formats didn’t exist. Finally, about two years ago, most of our agencies moved to NEMSIS-compliant ePCR software.
Our team understood early the advantage of this linkage, and we created an "XML importer" that would receive the data from the various ePCR vendors supporting our agencies. The importer opened the files and inserted the data into a SQL database, the framework of which is provided at the NEMSIS Web site.(1)
For the first time, our federation of cities became a "linked system" through our master ePCR database. We could analyze and compare care—apples to apples—across the region, and provide direct data support of our participation in various research trials. A cohesiveness emerged in our system that—prior to the work of such visionaries as Mears, who made the NEMSIS dream into reality—was not possible in the Dallas BioTel EMS System.
Our initial manual XML importer project was so successful that we have completed an automated XML importer upgrade. This updated software package runs on our secure server at UT Southwestern. It watches the assigned directories for the data to arrive by secure file transfer protocol (FTP) from each agency. When data appears in the directory, the automated XML importer sees the file, "un-zips" it, extracts the data, opens the master database, inserts the data, closes the database and closes the extracted file, archiving it for later use if necessary—all without human intervention, all on a 24/7 basis.
Dataflow improvements from our agency ePCR programs allow the BioTel system managers to:
The next version of NEMSIS—version 3.0—is expected to leverage the HL-7 messaging protocol used by most hospital IT systems. This will facilitate two-way data communication between EMS and the hospital. EMS patient data goes in and outcome data goes out.
On another front, a notable IMMEDIATE trial involves the direct transmission of ECGs from the monitors in the field to the national study office via the Internet. Further, ECG transmission for PCI lab activation has become a mainstay of many progressive EMS systems.
Finally, and perhaps most exciting, the era of the "ICU in the field" is likely upon us with data-streaming capabilities supported by such companies as General Devices and new work in pushing a significant volume of critical care data from field monitors to receiving hospitals by Philips Healthcare.
In my work in the Dallas-Fort Worth area with Dr. John Griswell, medical director of MedStar Ambulance Service, and his colleague Dr. Jeff Beeson, I have seen outstanding evidence of what a well-crafted monitoring system brings to prehospital emergency care. The sophistication of physical assessment, electronic communication with area emergency receiving facilities, and ongoing patient surveillance reaches a level comparable with virtually any intensive care unit in a hospital.
BioTel provides critical care support to about half of the ALS ambulance transports in the greater Dallas area. The call center is based in Parkland Hospital but provides protocol approval, consultation and receiving hospital communications for a large percentage of the calls involving stroke, trauma, respiratory distress and a combination of acute and chronically ill patients.
Most receiving hospitals in the U.S. do not have the resources to monitor these patients before they arrive. But BioTel and other pioneering hospitals are beginning to provide this service to begin treating patients before they arrive. This is appropriately blurring the lines between EMS and the hospital and more closely integrating out-of-hospital and in-hospital resources.
To effectively provide this service, call centers can’t rely on conversations between EMS responders and hospital providers. They need clinical patient data sent from the field before the patient arrives—and preferably during care and transport. Philips has been transmitting 12-lead ECGs on STEMI patients over the Internet to hospitals for the last four years. They are now extending this model for other critical care patients by sending vitals, ECG and Capnography waveforms while the ambulance is enroute to the hospital.
Prehospital transmission of patient data means that EMS is able to prepare the next level of care before arrival. Medical control centers or receiving hospitals can then begin triage, make space decisions, pull prior records, review the patient’s history and summon specialists (or decide specialists aren’t needed). This saves time, moves patients through the ED faster to other areas of the hospital, frees up valuable resources and expenditures, and expedites patient care and handoffs, returning EMS personnel to active status sooner.
It’s fitting that the acquisition of patient data and its organizational application should be taking such integral steps now, during this era in which "prehospital emergency care" is rapidly becoming a medical subspecialty. Perhaps years from now, the current era will be seen as a tipping point—when EMS and its astounding world of complex data requirements and providers joined forces with the House of Medicine in an overwhelming push to a bright future.
Disclosure: The author has reported no conflicts of interest with the sponsor of this supplement.